--- parser3/src/main/compile.y 2001/03/08 17:08:14 1.78 +++ parser3/src/main/compile.y 2001/07/25 10:33:05 1.151 @@ -1,39 +1,54 @@ -/* - $Id: compile.y,v 1.78 2001/03/08 17:08:14 paf Exp $ -*/ +/** @file + Parser: compiler(lexical parser and grammar). + + Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com) + Author: Alexander Petrosyan (http://design.ru/paf) +*/ %{ -#define YYSTYPE Array/**/ * +static char *RCSId="$Id: compile.y,v 1.151 2001/07/25 10:33:05 parser Exp $"; + +/** + @todo parser4: + - cache compiled code from request to request. to do that... + -#: make method definitions, @CLASS, @BASE, @USE instructions, + which would be executed afterwards, and actions + now performed at compile time would be delayed to run time. + -#: make cache expiration on time and on disk-change of class source + -#: in apache use subpools for compiled class storage + -#: in iis make up specialized Pool object for that +*/ + +#define YYSTYPE Array/**/ * #define YYPARSE_PARAM pc #define YYLEX_PARAM pc #define YYDEBUG 1 -#define YYERROR_VERBOSE +#define YYERROR_VERBOSE 1 #define yyerror(msg) real_yyerror((parse_control *)pc, msg) #define YYPRINT(file, type, value) yyprint(file, type, value) -#include -#include -#include - #include "compile_tools.h" #include "pa_value.h" #include "pa_request.h" #include "pa_vobject.h" #include "pa_vdouble.h" +#include "pa_globals.h" +#include "pa_vvoid.h" -#define SELF_NAME "self" -#define USE_NAME "USE" +#define SELF_ELEMENT_NAME "self" +#define USE_CONTROL_METHOD_NAME "USE" -int real_yyerror(parse_control *pc, char *s); +static int real_yyerror(parse_control *pc, char *s); static void yyprint(FILE *file, int type, YYSTYPE value); -int yylex(YYSTYPE *lvalp, void *pc); +static int yylex(YYSTYPE *lvalp, void *pc); // local convinient inplace typecast & var -#define PC ((parse_control *)pc) -#define POOL *PC->pool +#define PC (*(parse_control *)pc) +#define POOL (*PC.pool) #undef NEW #define NEW new(POOL) +#ifndef DOXYGEN %} %pure_parser @@ -43,6 +58,7 @@ int yylex(YYSTYPE *lvalp, void *pc); %token BOGUS %token BAD_STRING_COMPARISON_OPERATOR +%token BAD_HEX_LITERAL %token LAND "&&" %token LOR "||" @@ -63,20 +79,22 @@ int yylex(YYSTYPE *lvalp, void *pc); %token DEF "def" %token IN "in" %token FEXISTS "-f" +%token DEXISTS "-d" +%token IS "is" /* logical */ -%left "lt" "gt" "le" "ge" -%left "eq" "ne" -%left '<' '>' "<=" ">=" "##" -%left "==" "!=" +%left "##" %left "||" %left "&&" -%left "def" "in" "-f" +%left '<' '>' "<=" ">=" "lt" "gt" "le" "ge" +%left "==" "!=" "eq" "ne" +%left "is" "def" "in" "-f" "-d" %left '!' /* bitwise */ %left '#' -%left '&' '|' +%left '|' +%left '&' %left '~' /* numerical */ @@ -85,17 +103,15 @@ int yylex(YYSTYPE *lvalp, void *pc); %left NEG /* negation: unary - */ %% - -all: /* TODO: у ^execute непременно задать какой-то name, см. 'RUN' */ +all: one_big_piece { - String& MAIN=*NEW String(POOL); - MAIN.APPEND_CONST(MAIN_METHOD_NAME); Method& method=*NEW Method(POOL, - MAIN, - 0, /*numbered_params_count*/ + *main_method_name, + Method::CT_ANY, + 0, 0, /*min, max numbered_params_count*/ 0/*param_names*/, 0/*local_names*/, $1/*parser_code*/, 0/*native_code*/); - PC->vclass->add_method(MAIN, method); + PC.cclass->add_method(*main_method_name, method); } | methods; @@ -105,86 +121,87 @@ one_big_piece: maybe_codes; method: control_method | code_method; control_method: '@' STRING '\n' - control_strings { - String& command=*SLA2S($2); + maybe_control_strings { + const String& command=*LA2S($2); YYSTYPE strings_code=$4; if(strings_code->size()<1*2) { - strcpy(PC->error, "@"); - strcat(PC->error, command.cstr()); - strcat(PC->error, " is empty"); + strcpy(PC.error, "@"); + strcat(PC.error, command.cstr()); + strcat(PC.error, " is empty"); YYERROR; } if(command==CLASS_NAME) { - if(PC->vclass!=&PC->request->root_class) { // already changed from default? - strcpy(PC->error, "class already have a name '"); - strncat(PC->error, PC->vclass->name().cstr(), 100); - strcat(PC->error, "'"); + if(PC.cclass->base()) { // already changed from default? + strcpy(PC.error, "class already have a name '"); + strncat(PC.error, PC.cclass->name().cstr(), 100); + strcat(PC.error, "'"); YYERROR; } if(strings_code->size()==1*2) { // new class' name - String *name=SLA2S(strings_code); + const String *name=LA2S(strings_code); // creating the class - PC->vclass=NEW VClass(POOL); - PC->vclass->set_name(*name); - // defaulting base. may change with @BASE - PC->vclass->set_base(PC->request->root_class); + PC.cclass=NEW VClass(POOL); + PC.cclass->set_name(*name); // append to request's classes - PC->request->classes_array()+=PC->vclass; - PC->request->classes().put(*name, PC->vclass); + PC.request->classes().put(*name, PC.cclass); } else { - strcpy(PC->error, "@"CLASS_NAME" must contain sole name"); + strcpy(PC.error, "@"CLASS_NAME" must contain sole name"); YYERROR; } - } else { - if(command==USE_NAME) { - for(int i=0; isize(); i+=2) { - String *file=SLA2S(strings_code, i); - file->APPEND_CONST(".p"); - PC->request->use(file->cstr(), 0); - } - } else if(command==BASE_NAME) { - if(PC->vclass->base()!=&PC->request->root_class) { // already changed from default? - strcpy(PC->error, "there must be only one @"BASE_NAME); + } else if(command==USE_CONTROL_METHOD_NAME) { + for(int i=0; isize(); i+=2) + PC.request->use_file( + PC.request->absolute(*LA2S(strings_code, i))); + } else if(command==BASE_NAME) { + if(PC.cclass->base()) { // already changed from default? + strcpy(PC.error, "class already have a base '"); + strncat(PC.error, PC.cclass->base()->name().cstr(), 100); + strcat(PC.error, "'"); + YYERROR; + } + if(strings_code->size()==1*2) { + const String& base_name=*LA2S(strings_code); + VClass *base=static_cast( + PC.request->classes().get(base_name)); + if(!base) { + strcpy(PC.error, base_name.cstr()); + strcat(PC.error, ": undefined class in @"BASE_NAME); YYERROR; } - if(strings_code->size()==1*2) { - // TODO: преодолеть self и циклические base - String& base_name=*SLA2S(strings_code); - VClass *base=static_cast( - PC->request->classes().get(base_name)); - if(!base) { - strcpy(PC->error, base_name.cstr()); - strcat(PC->error, ": undefined class in @"BASE_NAME); - YYERROR; - } - PC->vclass->set_base(*base); - } else { - strcpy(PC->error, "@"BASE_NAME" must contain sole name"); + // @CLASS == @BASE sanity check + if(PC.cclass==base) { + strcpy(PC.error, "@"CLASS_NAME" equals @"BASE_NAME); YYERROR; } + PC.cclass->set_base(*base); } else { - strcpy(PC->error, command.cstr()); - strcat(PC->error, ": invalid special name. valid names are " - CLASS_NAME", "USE_NAME" and "BASE_NAME); + strcpy(PC.error, "@"BASE_NAME" must contain sole name"); YYERROR; } + } else { + strcpy(PC.error, "'"); + strncat(PC.error, command.cstr(), MAX_STRING/2); + strcat(PC.error, "' invalid special name. valid names are " + "'"CLASS_NAME"', '"USE_CONTROL_METHOD_NAME"' and '"BASE_NAME"'"); + YYERROR; } }; +maybe_control_strings: empty | control_strings; control_strings: control_string | control_strings control_string { $$=$1; P($$, $2) }; control_string: maybe_string '\n'; maybe_string: empty | STRING; code_method: '@' STRING bracketed_maybe_strings maybe_bracketed_strings maybe_comment '\n' maybe_codes { - const String *name=SLA2S($2); + const String *name=LA2S($2); YYSTYPE params_names_code=$3; Array *params_names=0; if(int size=params_names_code->size()) { params_names=NEW Array(POOL); for(int i=0; isize()) { locals_names=NEW Array(POOL); for(int i=0; ivclass->add_method(*name, method); + PC.cclass->add_method(*name, method); }; maybe_bracketed_strings: empty | bracketed_maybe_strings; @@ -214,18 +232,15 @@ maybe_comment: empty | STRING; maybe_codes: empty | codes; -codes: code | codes code { - $$=$1; - P($$, $2); -}; -code: write_str_literal | action; -action: get | put | with | call; +codes: code | codes code { $$=$1; P($$, $2) }; +code: write_string | action; +action: get | put | call; /* get */ get: get_value { $$=$1; /* stack: resulting value */ - O($$, OP_WRITE); /* value=pop; wcontext.write(value) */ + O($$, OP_WRITE_VALUE); /* value=pop; wcontext.write(value) */ }; get_value: '$' get_name_value { $$=$2 } get_name_value: name_without_curly_rdive EON | name_in_curly_rdive; @@ -237,8 +252,8 @@ name_without_curly_rdive: name_without_curly_rdive_read: name_without_curly_rdive_code { $$=N(POOL); Array *diving_code=$1; - String *first_name=SLA2S(diving_code); - if(first_name && *first_name==SELF_NAME) { + const String *first_name=LA2S(diving_code); + if(first_name && *first_name==SELF_ELEMENT_NAME) { O($$, OP_WITH_SELF); /* stack: starting context */ P($$, diving_code, /* skip over... */ @@ -259,10 +274,9 @@ name_without_curly_rdive_code: name_adva /* put */ -put: '$' name_expr_wdive constructor_value { +put: '$' name_expr_wdive construct { $$=$2; /* stack: context,name */ P($$, $3); /* stack: context,name,constructor_value */ - O($$, OP_CONSTRUCT); /* value=pop; name=pop; context=pop; construct(context,name,value) */ }; name_expr_wdive: name_expr_wdive_write @@ -271,8 +285,8 @@ name_expr_wdive: name_expr_wdive_write: name_expr_dive_code { $$=N(POOL); Array *diving_code=$1; - String *first_name=SLA2S(diving_code); - if(first_name && *first_name==SELF_NAME) { + const String *first_name=LA2S(diving_code); + if(first_name && *first_name==SELF_ELEMENT_NAME) { O($$, OP_WITH_SELF); /* stack: starting context */ P($$, diving_code, /* skip over... */ @@ -290,10 +304,29 @@ name_expr_wdive_root: ':' name_expr_dive }; name_expr_wdive_class: class_prefix name_expr_dive_code { $$=$1; P($$, $2) }; -constructor_value: - '[' any_constructor_code_value ']' { $$=$2 } -| '(' any_expr ')' { $$=$2 } +construct: + construct_square | + construct_round | + construct_curly ; +construct_square: '[' any_constructor_code_value ']' { + // stack: context, name + $$=$2; // stack: context, name, value + O($$, OP_CONSTRUCT_VALUE); /* value=pop; name=pop; context=pop; construct(context,name,value) */ +} +; +construct_round: '(' expr_value ')' { + // stack: context, name + $$=$2; // stack: context, name, value + O($$, OP_CONSTRUCT_EXPR); /* value=pop; name=pop; context=pop; construct(context,name,value) */ +} +; +construct_curly: '{' maybe_codes '}' { + // stack: context, name + $$=N(POOL); + CCA($$, $2); /* code=pop; name=pop; context=pop; construct(context,name,junction(code)) */ +}; + any_constructor_code_value: empty_string_value /* optimized $var[] case */ | STRING /* optimized $var[STRING] case */ @@ -312,13 +345,20 @@ codes__excluding_sole_str_literal: actio call: call_value { $$=$1; /* stack: value */ - O($$, OP_WRITE); /* value=pop; wcontext.write(value) */ + O($$, OP_WRITE_VALUE); /* value=pop; wcontext.write(value) */ }; call_value: '^' call_name store_params EON { /* ^field.$method{vasya} */ $$=$2; /* with_xxx,diving code; stack: context,method_junction */ O($$, OP_GET_METHOD_FRAME); /* stack: context,method_frame */ - P($$, $3); /* filling method_frame.store_params */ - O($$, OP_CALL); /* method_frame=pop; ncontext=pop; call(ncontext,method_frame) stack: value */ + + YYSTYPE params_code=$3; + if(params_code->size()==3) // probably [] case. [OP_VALUE + Void + STORE_PARAM] + if(Value *value=LA2V(params_code)) // it is OP_VALUE + value? + if(!value->is_defined()) // value is VVoid? + params_code=0; // ^zzz[] case. don't append lone empty param. + if(params_code) + P($$, params_code); // filling method_frame.store_params + O($$, OP_CALL); // method_frame=pop; ncontext=pop; call(ncontext,method_frame) stack: value }; call_name: name_without_curly_rdive; @@ -331,10 +371,7 @@ store_param: ; store_square_param: '[' store_code_param_parts ']' {$$=$2}; store_round_param: '(' store_expr_param_parts ')' {$$=$2}; -store_curly_param: '{' maybe_codes '}' { - $$=N(POOL); - PCA($$, $2); -}; +store_curly_param: '{' store_curly_param_parts '}' {$$=$2}; store_code_param_parts: store_code_param_part | store_code_param_parts ';' store_code_param_part { $$=$1; P($$, $3) } @@ -343,24 +380,30 @@ store_expr_param_parts: store_expr_param_part | store_expr_param_parts ';' store_expr_param_part { $$=$1; P($$, $3) } ; -store_code_param_part: - empty /* optimized [] case */ -| STRING { /* optimized [STRING] case */ - $$=$1; - O($$, OP_STORE_PARAM); -} -| constructor_code_value { /* [something complex] */ +store_curly_param_parts: + store_curly_param_part +| store_curly_param_parts ';' store_curly_param_part { $$=$1; P($$, $3) } +; +store_code_param_part: code_param_value { $$=$1; O($$, OP_STORE_PARAM); -} -; +}; store_expr_param_part: write_expr_value { $$=N(POOL); + PEA($$, $1); +}; +store_curly_param_part: maybe_codes { + $$=N(POOL); PCA($$, $1); }; -write_expr_value: any_expr { +code_param_value: + empty_void_value /* optimized [;...] case */ +| STRING /* optimized [STRING] case */ +| constructor_code_value /* [something complex] */ +; +write_expr_value: expr_value { $$=$1; - O($$, OP_WRITE); + O($$, OP_WRITE_EXPR_RESULT); }; /* name */ @@ -394,7 +437,7 @@ name_expr_with_subvar_value: STRING subv $$=N(POOL); O($$, OP_CREATE_EWPOOL); P($$, $1); - O($$, OP_WRITE); + O($$, OP_WRITE_VALUE); P($$, $2); O($$, OP_REDUCE_EWPOOL); }; @@ -421,32 +464,18 @@ class_prefix: STRING ':' { }; -/* with */ - -with: '$' name_without_curly_rdive '{' codes '}' { - $$=$2; - O($$, OP_CREATE_RWPOOL); - P($$, $4); - O($$, OP_REDUCE_RWPOOL); - O($$, OP_WRITE); -}; - /* expr */ -any_expr: - empty_double_value /* optimized $var() case */ -| optimized_expr /* $var(something) */ -; -optimized_expr: expr { - if(($$=$1)->size()==2) { // only one string literal in there? +expr_value: expr { + if(($$=$1)->size()==2) // only one string literal in there? change_string_literal_to_double_literal($$); // make that string literal Double - } }; expr: STRING | get_value | call_value | '"' string_inside_quotes_value '"' { $$ = $2; } +| '\'' string_inside_quotes_value '\'' { $$ = $2; } | '(' expr ')' { $$ = $2; } /* stack: operand // stack: @operand */ | '-' expr %prec NEG { $$=$2; O($$, OP_NEG) } @@ -455,6 +484,7 @@ expr: | "def" expr { $$=$2; O($$, OP_DEF) } | "in" expr { $$=$2; O($$, OP_IN) } | "-f" expr { $$=$2; O($$, OP_FEXISTS) } +| "-d" expr { $$=$2; O($$, OP_DEXISTS) } /* stack: a,b // stack: a@b */ | expr '-' expr { $$=$1; P($$, $3); O($$, OP_SUB) } | expr '+' expr { $$=$1; P($$, $3); O($$, OP_ADD) } @@ -479,6 +509,7 @@ expr: | expr "ge" expr { $$=$1; P($$, $3); O($$, OP_STR_GE) } | expr "eq" expr { $$=$1; P($$, $3); O($$, OP_STR_EQ) } | expr "ne" expr { $$=$1; P($$, $3); O($$, OP_STR_NE) } +| expr "is" expr { $$=$1; P($$, $3); O($$, OP_IS) } ; string_inside_quotes_value: maybe_codes { @@ -490,16 +521,17 @@ string_inside_quotes_value: maybe_codes /* basics */ -write_str_literal: STRING { - $$=$1; - O($$, OP_WRITE); +write_string: STRING { + // optimized from OP_STRING+OP_WRITE_VALUE to OP_STRING__WRITE + change_string_literal_to_write_string_literal($$=$1) }; -empty_double_value: /* empty */ { $$=VL(NEW VDouble(*NEW Double(POOL, 0))) }; +empty_void_value: /* empty */ { $$=VL(NEW VVoid(POOL)) }; empty_string_value: /* empty */ { $$=VL(NEW VString(POOL)) }; empty: /* empty */ { $$=N(POOL) }; %% +#endif /* 000$111(2222)00 @@ -517,54 +549,90 @@ empty: /* empty */ { $$=N(POOL) }; 4:[^({]=pop */ -int yylex(YYSTYPE *lvalp, void *pc) { - #define lexical_brackets_nestage PC->brackets_nestages[PC->sp] +static int yylex(YYSTYPE *lvalp, void *pc) { + #define lexical_brackets_nestage PC.brackets_nestages[PC.sp] #define RC {result=c; goto break2; } register int c; int result; - if(PC->pending_state) { - result=PC->pending_state; - PC->pending_state=0; + if(PC.pending_state) { + result=PC.pending_state; + PC.pending_state=0; return result; } - char *begin=PC->source; - char *end; - int begin_line=PC->line; + const char *begin=PC.source; + const char *end; + int begin_line=PC.line; int skip_analized=0; while(true) { - c=*(end=(PC->source++)); + c=*(end=(PC.source++)); if(c=='\n') { - PC->line++; - PC->col=0; + PC.line++; + PC.col=0; } else - PC->col++; - - // todo: # in 0+1 column comment + PC.col++; - // escaping: ^^ ^$ ^; ^) ^} ^( ^{ ^" - if(c=='^') - switch(*PC->source) { + if(c=='^' && PC.ls!=LS_COMMENT && PC.ls!=LS_DEF_COMMENT) + switch(*PC.source) { + // escaping: ^^ & co case '^': case '$': case ';': + case '(': case ')': case '[': case ']': case '{': case '}': - case '"': + case '"': if(end!=begin) { // append piece till ^ - PC->string->APPEND(begin, end-begin, PC->file, begin_line); + PC.string->APPEND_CLEAN(begin, end-begin, PC.file, begin_line); } // reset piece 'begin' position & line - begin=PC->source; // ^ - begin_line=PC->line; + begin=PC.source; // ^ + begin_line=PC.line; // skip over ^ and _ - PC->source++; PC->col++; + PC.source++; PC.col++; // skip analysis = forced literal continue; + + // converting ^#HH into char(hex(HH)) + case '#': + if(end!=begin) { + // append piece till ^ + PC.string->APPEND_CLEAN(begin, end-begin, PC.file, begin_line); + } + // #HH ? + if(PC.source[0]=='#' && PC.source[1] && PC.source[2]) { + char *hex=(char *)POOL.malloc(1); + hex[0]= + hex_value[(unsigned char)PC.source[1]]*0x10+ + hex_value[(unsigned char)PC.source[2]]; + if(hex[0]==0) { + result=BAD_HEX_LITERAL; + goto break2; // wrong hex value[no ^#00 chars allowed]: bail out + } + // append char(hex(HH)) + PC.string->APPEND_CLEAN(hex, 1, PC.file, begin_line); + // skip over ^#HH + PC.source+=3; + PC.col+=3; + // reset piece 'begin' position & line + begin=PC.source; // ^ + begin_line=PC.line; + continue; + } + break; + } + // #comment start skipping + if(c=='#' && PC.col==1) { + if(end!=begin) { + // append piece till # + PC.string->APPEND_CLEAN(begin, end-begin, PC.file, begin_line); } - switch(PC->ls) { + // fall into COMMENT lexical state [wait for \n] + push_LS(PC, LS_COMMENT); + } + switch(PC.ls) { // USER'S = NOT OURS case LS_USER: @@ -576,7 +644,7 @@ int yylex(YYSTYPE *lvalp, void *pc) { push_LS(PC, LS_METHOD_NAME); RC; case '@': - if(PC->col==0+1) { + if(PC.col==0+1) { push_LS(PC, LS_DEF_NAME); RC; } @@ -584,12 +652,31 @@ int yylex(YYSTYPE *lvalp, void *pc) { } break; + // #COMMENT + case LS_COMMENT: + if(c=='\n') { + // skip comment + begin=PC.source; + begin_line=PC.line; + + pop_LS(PC); + continue; + } + break; + // STRING IN EXPRESSION - case LS_EXPRESSION_STRING: + case LS_EXPRESSION_STRING_QUOTED: + case LS_EXPRESSION_STRING_APOSTROFED: switch(c) { case '"': - pop_LS(PC); //"abc". - RC; + case '\'': + if( + PC.ls == LS_EXPRESSION_STRING_QUOTED && c=='"' || + PC.ls == LS_EXPRESSION_STRING_APOSTROFED && c=='\'') { + pop_LS(PC); //"abc". | 'abc'. + RC; + } + break; case '$': push_LS(PC, LS_VAR_NAME_SIMPLE); RC; @@ -603,10 +690,10 @@ int yylex(YYSTYPE *lvalp, void *pc) { case LS_DEF_NAME: switch(c) { case '[': - PC->ls=LS_DEF_PARAMS; + PC.ls=LS_DEF_PARAMS; RC; case '\n': - PC->ls=LS_DEF_SPECIAL_BODY; + PC.ls=LS_DEF_SPECIAL_BODY; RC; } break; @@ -616,7 +703,7 @@ int yylex(YYSTYPE *lvalp, void *pc) { case ';': RC; case ']': - PC->ls=*PC->source=='['?LS_DEF_LOCALS:LS_DEF_COMMENT; + PC.ls=*PC.source=='['?LS_DEF_LOCALS:LS_DEF_COMMENT; RC; case '\n': // wrong. bailing out pop_LS(PC); @@ -630,7 +717,7 @@ int yylex(YYSTYPE *lvalp, void *pc) { case ';': RC; case ']': - PC->ls=LS_DEF_COMMENT; + PC.ls=LS_DEF_COMMENT; RC; case '\n': // wrong. bailing out pop_LS(PC); @@ -646,8 +733,13 @@ int yylex(YYSTYPE *lvalp, void *pc) { break; case LS_DEF_SPECIAL_BODY: + // @todo in case + // ################ + // @next-method + // we are here with c=='@' + // which is wrong, and need action if(c=='\n') { - switch(*PC->source) { + switch(*PC.source) { case '@': case 0: // end of special_code pop_LS(PC); break; @@ -662,9 +754,9 @@ int yylex(YYSTYPE *lvalp, void *pc) { switch(c) { case ')': if(--lexical_brackets_nestage==0) - if(PC->ls==LS_METHOD_ROUND) // method round param ended - PC->ls=LS_METHOD_AFTER; // look for method end - else // PC->ls==LS_VAR_ROUND // variable constructor ended + if(PC.ls==LS_METHOD_ROUND) // method round param ended + PC.ls=LS_METHOD_AFTER; // look for method end + else // PC.ls==LS_VAR_ROUND // variable constructor ended pop_LS(PC); // return to normal life RC; case '$': @@ -677,25 +769,33 @@ int yylex(YYSTYPE *lvalp, void *pc) { lexical_brackets_nestage++; RC; case '-': - if(*PC->source=='f') { // -f + switch(*PC.source) { + case 'f': // -f skip_analized=1; result=FEXISTS; - } else + goto break2; + case 'd': // -d + skip_analized=1; + result=DEXISTS; + goto break2; + default: result=c; + goto break2; + } goto break2; case '+': case '*': case '/': case '%': case '~': case ';': RC; case '&': case '|': case '#': - if(*PC->source==c) { // && || + if(*PC.source==c) { // && || result=c=='#'?LXOR:c=='&'?LAND:LOR; skip_analized=1; } else result=c; goto break2; case '<': case '>': case '=': case '!': - if(*PC->source=='=') { // <= >= == != + if(*PC.source=='=') { // <= >= == != skip_analized=1; switch(c) { case '<': result=NLE; break; @@ -707,37 +807,49 @@ int yylex(YYSTYPE *lvalp, void *pc) { result=c; goto break2; case '"': - push_LS(PC, LS_EXPRESSION_STRING); + push_LS(PC, LS_EXPRESSION_STRING_QUOTED); + RC; + case '\'': + push_LS(PC, LS_EXPRESSION_STRING_APOSTROFED); RC; case 'l': case 'g': case 'e': case 'n': if(end==begin) // right after whitespace - switch(*PC->source) { -// case '?': // ok [and bad cases, yacc would bark at them] - case 't': // lt gt [et nt] - result=c=='l'?SLT:c=='g'?SGT:BAD_STRING_COMPARISON_OPERATOR; - skip_analized=1; - goto break2; - case 'e': // le ge ne [ee] - result=c=='l'?SLE:c=='g'?SGE:c=='n'?SNE:BAD_STRING_COMPARISON_OPERATOR; - skip_analized=1; - goto break2; - case 'q': // eq [lq gq nq] - result=c=='e'?SEQ:BAD_STRING_COMPARISON_OPERATOR; - skip_analized=1; - goto break2; + if(isspace(PC.source[1])) { + switch(*PC.source) { + // case '?': // ok [and bad cases, yacc would bark at them] + case 't': // lt gt [et nt] + result=c=='l'?SLT:c=='g'?SGT:BAD_STRING_COMPARISON_OPERATOR; + skip_analized=1; + goto break2; + case 'e': // le ge ne [ee] + result=c=='l'?SLE:c=='g'?SGE:c=='n'?SNE:BAD_STRING_COMPARISON_OPERATOR; + skip_analized=1; + goto break2; + case 'q': // eq [lq gq nq] + result=c=='e'?SEQ:BAD_STRING_COMPARISON_OPERATOR; + skip_analized=1; + goto break2; + } } break; case 'i': if(end==begin) // right after whitespace - if(PC->source[0]=='n') { // in - skip_analized=1; - result=IN; - goto break2; + if(isspace(PC.source[1])) { + switch(PC.source[0]) { + case 'n': // in + skip_analized=1; + result=IN; + goto break2; + case 's': // is + skip_analized=1; + result=IS; + goto break2; + } } break; case 'd': if(end==begin) // right after whitespace - if(PC->source[0]=='e' && PC->source[1]=='f') { // def + if(PC.source[0]=='e' && PC.source[1]=='f') { // def skip_analized=2; result=DEF; goto break2; @@ -751,8 +863,8 @@ int yylex(YYSTYPE *lvalp, void *pc) { // that's a leading|traling space or after-operator-space // ignoring it // reset piece 'begin' position & line - begin=PC->source; // after whitespace char - begin_line=PC->line; + begin=PC.source; // after whitespace char + begin_line=PC.line; continue; } break; @@ -760,14 +872,23 @@ int yylex(YYSTYPE *lvalp, void *pc) { // VARIABLE GET/PUT/WITH case LS_VAR_NAME_SIMPLE: case LS_EXPRESSION_VAR_NAME: - if(PC->ls==LS_EXPRESSION_VAR_NAME) { + case LS_VAR_NAME_NO_COLON: + if(PC.ls==LS_EXPRESSION_VAR_NAME) { // name in expr ends also before binary operators switch(c) { - case '+': case '-': case '*': case '/': case '%': - case '&': case '|': - case '<': case '>': case '=': case '!': + case '-': + pop_LS(PC); + PC.source--; if(--PC.col<0) { PC.line--; PC.col=-1; } + result=EON; + goto break2; + } + } + if(PC.ls==LS_VAR_NAME_NO_COLON) { + // name already has ':', stop before next + switch(c) { + case ':': pop_LS(PC); - PC->source--; if(--PC->col<0) { PC->line--; PC->col=-1; } + PC.source--; if(--PC.col<0) { PC.line--; PC.col=-1; } result=EON; goto break2; } @@ -776,31 +897,41 @@ int yylex(YYSTYPE *lvalp, void *pc) { case 0: case ' ': case '\t': case '\n': case ';': - case ']': case '}': case ')': case '"': + case ']': case '}': case ')': + case '"': case '\'': + case '<': case '>': // these stand for HTML brackets AND expression binary ops + case '+': case '*': case '/': case '%': + case '&': case '|': + case '=': case '!': + // common delimiters + case ',': + // before call + case '^': pop_LS(PC); - PC->source--; if(--PC->col<0) { PC->line--; PC->col=-1; } + PC.source--; if(--PC.col<0) { PC.line--; PC.col=-1; } result=EON; goto break2; case '[': - PC->ls=LS_VAR_SQUARE; + PC.ls=LS_VAR_SQUARE; lexical_brackets_nestage=1; RC; case '{': if(begin==end) { // ${name}, no need of EON, switching LS - PC->ls=LS_VAR_NAME_CURLY; + PC.ls=LS_VAR_NAME_CURLY; } else { - PC->ls=LS_VAR_CURLY; + PC.ls=LS_VAR_CURLY; lexical_brackets_nestage=1; } RC; case '(': - PC->ls=LS_VAR_ROUND; + PC.ls=LS_VAR_ROUND; lexical_brackets_nestage=1; RC; case '.': // name part delim case '$': // name part subvar case ':': // ':name' or 'class:name' + PC.ls=LS_VAR_NAME_NO_COLON; // stop before next ':' RC; } break; @@ -863,15 +994,15 @@ int yylex(YYSTYPE *lvalp, void *pc) { case LS_METHOD_NAME: switch(c) { case '[': - PC->ls=LS_METHOD_SQUARE; + PC.ls=LS_METHOD_SQUARE; lexical_brackets_nestage=1; RC; case '{': - PC->ls=LS_METHOD_CURLY; + PC.ls=LS_METHOD_CURLY; lexical_brackets_nestage=1; RC; case '(': - PC->ls=LS_METHOD_ROUND; + PC.ls=LS_METHOD_ROUND; lexical_brackets_nestage=1; RC; case '.': // name part delim @@ -893,7 +1024,7 @@ int yylex(YYSTYPE *lvalp, void *pc) { RC; case ']': if(--lexical_brackets_nestage==0) { - PC->ls=LS_METHOD_AFTER; + PC.ls=LS_METHOD_AFTER; RC; } break; @@ -911,9 +1042,11 @@ int yylex(YYSTYPE *lvalp, void *pc) { case '^': push_LS(PC, LS_METHOD_NAME); RC; + case ';': // param delim + RC; case '}': if(--lexical_brackets_nestage==0) { - PC->ls=LS_METHOD_AFTER; + PC.ls=LS_METHOD_AFTER; RC; } break; @@ -925,22 +1058,22 @@ int yylex(YYSTYPE *lvalp, void *pc) { case LS_METHOD_AFTER: if(c=='[') {/* ][ }[ )[ */ - PC->ls=LS_METHOD_SQUARE; + PC.ls=LS_METHOD_SQUARE; lexical_brackets_nestage=1; RC; } if(c=='{') {/* ]{ }{ ){ */ - PC->ls=LS_METHOD_CURLY; + PC.ls=LS_METHOD_CURLY; lexical_brackets_nestage=1; RC; } if(c=='(') {/* ]( }( )( */ - PC->ls=LS_METHOD_ROUND; + PC.ls=LS_METHOD_ROUND; lexical_brackets_nestage=1; RC; } pop_LS(PC); - PC->source--; if(--PC->col<0) { PC->line--; PC->col=-1; } + PC.source--; if(--PC.col<0) { PC.line--; PC.col=-1; } result=EON; goto break2; } @@ -955,41 +1088,34 @@ break2: if((c=='@' || c==0) && end[-1]=='\n') { // we are before LS_DEF_NAME or EOF? // strip last \n end--; + if(end!=begin && end[-1]=='\n') // allow one empty line before LS_DEF_NAME + end--; } - if(end!=begin) { // last piece still alive? + if(end!=begin && PC.ls!=LS_COMMENT) { // last piece still alive and not comment? // append it - PC->string->APPEND(begin, end-begin, PC->file, begin_line/*, start_col*/); + PC.string->APPEND_CLEAN(begin, end-begin, PC.file, begin_line/*, start_col*/); } } - if(PC->string->size()) { // something accumulated? + if(PC.string->size()) { // something accumulated? // create STRING value: array of OP_VALUE+vstring - *lvalp=VL(NEW VString(*PC->string)); + *lvalp=VL(NEW VString(*PC.string)); // new pieces storage - PC->string=NEW String(POOL); + PC.string=NEW String(POOL); // make current result be pending for next call, return STRING for now - PC->pending_state=result; result=STRING; + PC.pending_state=result; result=STRING; } if(skip_analized) { - PC->source+=skip_analized; PC->col+=skip_analized; + PC.source+=skip_analized; PC.col+=skip_analized; } return result; } -int real_yyerror(parse_control *pc, char *s) /* Called by yyparse on error */ - { - //fprintf(stderr, "[%s]\n", s); - - strncpy(pc->error, s, MAX_STRING); // TODO: перепроверить с треклятым последним байтом +static int real_yyerror(parse_control *pc, char *s) { // Called by yyparse on error + strncpy(PC.error, s, MAX_STRING); return 1; - } - -static void - yyprint( - FILE *file, - int type, - YYSTYPE value) - { - if(type==STRING) - fprintf(file, " \"%s\"", SLA2S(value)->cstr()); - } +} +static void yyprint(FILE *file, int type, YYSTYPE value) { + if(type==STRING) + fprintf(file, " \"%s\"", LA2S(value)->cstr()); +}